基于Wasserstein距离的多电压暂降事件同源检测方法  

Multiple Voltage Sag Events Homology Detection Based on Wasserstein Distance

作　　者：肖先勇[1] 桂良宇 李成鑫[1] 张华赢 李鸿鑫 汪清 XIAO Xianyong;GUI Liangyu;LI Chengxin;ZHANG Huaying;LI Hongxin;WANG Qing(College of Electrical Engineering,Sichuan University,Chengdu 610065,Sichuan Province,China;Shenzhen Power Supply Co.,Ltd.,Shenzhen 518000,Guangdong Province,China)

机构地区：[1]四川大学电气工程学院,四川省成都市610065 [2]深圳供电局有限公司,广东省深圳市518000

出　　处：《电网技术》2020年第12期4684-4693,共10页Power System Technology

基　　金：国家自然科学基金项目(51807126);中国南方电网有限责任公司科技项目(090000KK52180110)的资助。

摘　　要：一次电网故障会触发多个变电站的电能质量监测装置,记录多条电压暂降事件。同源多数据的冗余信息,极大地增加了电压暂降监测数据处理的计算量和复杂度。提出多电压暂降事件同源检测概念和基于Wasserstein距离的多电压暂降事件同源检测方法。提出电压暂降经单台或多台变压器的传播规律,基于Wasserstein距离度量,构建可能性矩阵,度量多电压暂降波形数据间的相似程度。在缺失电压暂降源数量信息的前提下,采用基于密度的DBSCAN聚类方法对度量结果进行聚类分析,得到多电压暂降事件同源检测结果。在PASCD/EMTDC平台搭建IEEE14节点仿真,生成多电压暂降事件,应用仿真数据和广东电网某市实测数据验证所提方法。结果表明所提方法能够准确实现多电压暂降事件的同源检测,并对实测数据中的坏数据、仅被一个监测装置记录的事件具有一定的识别效果,证明了所提方法的可行性与正确性。A short-circuit fault in a power grid will trigger the power quality(PQ) monitoring devices which are fixed in multiple substations, and the PQ system records multiple voltage sag events. The redundant information of multi-data from the same voltage sag source increases the calculation workload and complexity of data processing greatly. This paper proposes the conception of multiple voltage sag events homology detection(MSHD), and presents the MSHD method based on the Wasserstein distance. This paper first discusses the voltage sag propagation laws through one or more transformers;then a possibility matrix is built based on Wasserstein distance measurement to measure the similarity between the multiple voltage sag waveform data. The measurement results are clustered using DBSCAN clustering method based on density to get the MSHD result, when there is not enough information of the number of sag sources. Finally an IEEE 14-bus test system on PSCAD/EMTDC platform is constructed to generate the voltage sag waveform. The accuracy of the proposed method is verified using the simulation data and measured data from a city in Guangdong power grid. The results show that the proposed method can do MSHD accurately, and it can detect the bad data among the measured voltage sag data and the single event recorded by only one PQ monitor, which proves the feasibility and correctness of this method.

关 键 词：电压暂降 多事件同源检测 Wasserstein距离 监测装置 DBSCAN聚类 

分 类 号：TM721[电气工程—电力系统及自动化]